3138
EUWARU C. TAYLOR AND WENUELLA. EHRHART [CONTRIBUTION FROM THE
FRICK C H E M I C A L LABORATORY, PRINCETON
Vol. 82
~SIVERSITY]
A Convenient Synthesis of Formamidine and Acetamidine Acetate BY EDWARD C. TAYLOR AND WENDELL -1. EHRHART~ RECEIVED SOVEMBER 10,1959 It is shown t h a t t h e reaction of ethyl orthoformate, ammonia and 11/1moles of acetic acid gives pure formamidine acetate in 8470 yield. Acetamidine acetate m a p also be prepared in 845; 1-ield by the reaction of ethyl orthoacetate with ammonia and ammonium acetate. I n contrast t o formamidine and acetamidine hydrochloride, which are highly deliquescent, the corresponding acetates are not appreciably hygroscopic. Formamidine acetate may be used directly in many condensation reactions without prior liberation of free formamidine. Thus, 4-aminopyrimido [4,5-d]-pyrimidine ( I V ) m a y be prepared by reaction of formamidine acetate with (a) 1/3 mole of malononitrile, ( b ) 4-amino-5-cyanopyrimidine, (c) 4-methylmercapto5-formyl-6-aminopyrimidine, ( d ) 4-aminopyrimidine-5thiocarboxamideor ( e ) 4-methylmercapto- or 4-mercaptopyrimido[4,5-d]pyrimidine. Direct condensation of formamidine acetate with ethyl cyanoacetate yields ethyl aminomethylenecyanoacetate, while reaction with o-phenylenediamine gives benzimidazole. Reaction of formamidine acetate with acetic anhydride yields triacetylaminomethane (VI), identical with a compound previously prepared b y Pinner from formamidine hydrochloride, acetic anhydride and sodium acetate b u t recently considered t o be diacetylformamidine (V).
Previous work in this Laboratory has shown that various o-aminonitriles in the pyrazole2 and imidazole2,3series may readily be converted to fused pyrimidine heterocycles by reaction with ethyl orthoformate and acetic anhydride to give an ethoxymethyleneamino derivative I , which upon subsequent reaction with ammonia or primary amines undergoes cyclization to 11. In connection with current work on the chemistry and properties of pyrimido [4,5-d]pyrimidines, we attempted a similar series of reactions with some 4-amino-5cyanopyrimidines, and the unexpected results obtained led to a new synthesis of formamidine which is described in the present paper.
NH
Following the procedures previously employed with other o-aminonitriles, a mixture of 4-amino-5cyanopyrimidine (III), ethyl orthoformate and acetic anhydride was heated under reflux and then evaporated under reduced pressure to a viscous oil. Since this presumed ethoxymethylene derivative could not be crystallized, it was treated directly with ethanolic ammonia. T o our surprise, unchanged I11 separated in quantitative yield from the reaction mixture. 4-Amino-5-cyano-6-methylpyrimidine similarly was recovered unchanged when treated under the same conditions. Since monosubstituted pyrimido [4,5-d]pyrimidines are known to be labile to aqueous base4 and imino ethers such as (I) are hydrolyzed under the same conditions, it was thought that the failure of this reaction might have been due to the presence of moisture in the alcoholic ammonia. A4ccordingly, 4-amino-5-cyanopyrimidine (111) was heated with an equimolar mixture of acetic anhydride and ethyl orthoforinate while a stream of dry ammonia was bubbled directly through the reaction mixture. Filtration yielded the desired 4aminopyrimido[4,5-d]pyrimidine (IV) and a large amount of white water-soluble crystals. I t was evident from the bulk of this latter material that it was not a derivative of 4-arnino-5-cyanopyrimiditie; indeed, (1 j Parke, Davis and Co. Fellow in Chemistry, 1958-1959. (2) E. C. Taylor and K . S . Hartke, THIS J O U R N A L , 81, 2459 (1959). (3) E. C. Taylor and P . K. Loeffler, ibid.,82, 3147 (1960). (4) K . J. Knopf, P h D. Thesis, Princeton University, 1957.
it could be prepared in equivalent amount by bubbling ammonia directly into a refluxing solution of ethyl orthoformate in acetic anhydride. Fractional crystallization of this material from ethanol yielded a less soluble fraction (A), m.p. 279-280" dec., and a soluble fraction (B), m.p. 162-1 64", as the major constituent. Compound B, C3H8N202, was shown to be formamidine acetate by conversion to formamidine hydrochloride by means of dry hydrogen chloride, to formamidine picrate with picric acid and by its utilization in synthesis (vide injm). Compound A proved to be identical witll a compound previously prepared by Pinner5 by the reaction of acetic anhydride with an equimolar mixture of formamidine hydrochloride and sodium acetate. Although it was first claimed to be diacetylformamidine (V), it was later stated6that the analytical results were better in accord with the structure triacetylaminomethane (VI), although no evidence in support of this assignment could be advanced. An acetyl determination on compound A now definitely confirms this structure.' I t should be pointed out that triacetylaminomethane (VI) can be prepared directly from formamidine acetate and acetic anhydride and this procedure is preferable to Pinner's original method. Compound VI has recently been prepared by a different procedure by Bredereck, HC(=XCOCHB)SHCOCHa 5'
HC( S H C O C H I ) ~ VI
et a1.,8but no physical properties for this substance were given and its relationship to Pinner's compound was not mentioned. A recent review article on amidinesg assigns structure V to Pinner's compound, apparently unaware of his later structural revision. The formation of formamidine acetate by the reaction of acetic anhydride, ethyl orthoformate and ammonia is of great interest, for we believe that it represents the first synthesis of an unsubsti(5) A. Pinner, Bet., 16, 1659 (1883); 17, 171 (1884). (6) A . Pinner, "Die Imidoather und ihre Derivate," Berlin, 1692. (7) T h e acetyl determination was carried out by the method described by Freudenberg ( A n n . , 433, 230 (1923)) and indicated four acetyl groups. However, i t has been shown (ref. 12) t h a t triformylaminomethane gives four formyl groups by this method and i t seems reasonable t o assume t h a t formic acid would also be produced from V I , thus accounting for the fourth acetyl group. (8) H. Bredereck, R . Gompper, H. Rempfer, H. Keck and K. Klemm. ATiEew. C h e i i i . , 70, 2F9 (1958). (9) I
